The Study Of Noninvasive Methods In Testing Cardiovascular Autonomic Neuropathy

Objective:Diabetic autonomic neuropathy is one of the most common and serious complications of diabetes. Cardiovascular autonomic neuropathy (CAN) is the most studied and clinically important form among its various and variable performances. CAN, which may lead to serious consequences such as increased occurrence of cardiovascular events, higher risk of sudden death, decreased survival ratio and poor quality of life in diabetic patients, may occur in early age of diabetes . The clinical manifestation of CAN can be performed as concealed, lacking of specificity. The reported prevalence varies widely from 7. 7% to 90% depending on the cohort studied and the methods of assessment. It showed that the 5-year mortality rate of this serious complication is apparently higher in diabetic people with CAN than those without cardiovascular autonomic involved.This research assesses cardiovascular autonomic function in healthy controls and people with diabetes according the methods of power spectral analysis of heat rate variability (HRV), baroreflex sensitivity (BRS) and blood pressure variability (BPV). Thus we can explore the characteristics of power spectral analysis in diabetes. We also make the multivariate linear regression analysis of each power spectral index to conclude the influence factors. Furthermore, we assessed cardiovascular autonomic neuropathy by the five autonomic functions of Ewing's methodology and draw the ROC curve of each power spectral index. We aims to provide a group of convenient, available and noninvasive methods to evaluate CAN in diabetes.Methods:1. There were 118 patients of type 2 diabetic mellitus in Huashan Hospital, Department of Endocrinology, and 118 healthy controls from community and doctors worked in our hospital(January 2006 to January 2008).2. 5-min power spectral analysis of HRV, BRS and BPV were used to estimate CAN.3. The characteristics of CAN were compared between 118 diabetic patients and 118 healthy controls.4. Multivariate linear regression analysis was made to evaluate the influence factors of each power spectral index of HRV, BRS and BPV.5. 53 diabetic patients and 35 healthy controls also underwent the five cardiovascular reflex tests (including heat rate variation during Valsalva maneuver and deep breathing, 30:15 ratio, blood pressure response to standing and to sustained handgrip). According to the result of the five tests , ROC curve was made to decide the cut point of each power spectral index and assess its value in diagnosing CAN.6. All statistical analyses were performed using SPSS 11.5 statistical analysis.Results:1. The resting heat rate is faster in diabetic patients than healthy cohorts. The absolute value of total variability (TV), very low frequency (VLF), low frequency (LF) and high frequency (HF) of HRV decreased in diabetic group compared with the control group(P<0.05, P<0.01或P<0.001) , though the normalization value of VLF, LF, HF of HRV in diabetic group has no difference between the two groups. All the power spectral indexes of BRS have no significant difference between the diabetes group and the control group. The absolute value of TV, VLF of BPV increased in diabetic group (P<0.05或P<0.01) .2. Each index of HRV, BRS and BPV has nothing to do with the sex. The resting heart rate and the history of diabetes are positively correlated, while the RRI mean is negatively correlated with diabetes. The TV, LF and HF ingredients of HRV, which are negatively correlated with age and diabetes history, are positively correlated with BMI and have no correlation with hypertension. The index of BRS, negatively correlated with the systolic pressure level, has nothing to do with diabetes and hypertension. The power spectral indexes of BPV are positively correlated with hypertension and the systolic pressure level. They are not correlated with diabetes except VLF.3. The TV, the absolute value of VLF, LF and HF all decreased as age increased in both diabetic and control group.4. The five cardiovascular reflex tests were down to detect CAN. Then ROC curve was draw to determine the diagnostic utility of each power spectral index of HRV, BRS and BPV. The results indicate that the ROC area of the spectrum components of HRV and BRS under the curve varies from 0.63 to 0.71, while the sensitivity and specificity varies from 0. 55 to 0. 72. It also indicates that the spectrum components of BPV have poor consistency with the five cardiovascular reflex tests.Conclusions:1. Diabetic patients have quick and fixed heart rates. The power spectral indexes of HRV decreased and BPV increased in people with diabetes.2. The power spectral indexes of HRV decreased as age grow. They also increased with the increasing of body mass index .3. Our multivariate linear regression shows that the power spectral indexes of BRS were negatively correlated with the systolic pressure level and have no significant correlation with diabetes and hypertension.4. The power spectral indexes of BPV are positively correlated with hypertension and the systolic pressure level. They are not correlated with diabetes. 5. The power spectral analysis of HRV and BRS can be used in diagnosis of CAN. In order to be widely used in estimating CAN, further study of the spectrum analysis of BPV is needed.6. Older, longer sufferance of diabetes mellitus and lower body mass index go together with decreased HRV in diabetic patients.